Stockfish Testing Queue

Finished - 24198 tests

20-02-18 pr ps_leverpush diff
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 56669 W: 12582 L: 12530 D: 31557
sprt @ 10+0.1 th 1 simplification: leverPush doesn't seem to do much on my local machines.
22-02-18 sn gradient diff
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 26352 W: 5938 L: 5676 D: 14738
sprt @ 10+0.1 th 1 Adjust gradient based on current bestValue
21-02-18 jd ppBlocked diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 12345 W: 2695 L: 2766 D: 6884
sprt @ 10+0.1 th 1 Give a small bonus even if our passed pawn is blocked if we attacked the blocking piece and it is not defended.
21-02-18 fa RT diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 41069 W: 9017 L: 9019 D: 23033
sprt @ 10+0.1 th 1 Rook Safe Check +15
21-02-18 sn LSB2'' diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 10493 W: 2249 L: 2329 D: 5915
sprt @ 10+0.1 th 1 Try k+=2. Bench: 5619406
21-02-18 SC fb9f7abc369c1baa0b95867 diff
ELO: 166.98 +-3.0 (95%) LOS: 100.0%
Total: 30000 W: 15430 L: 2028 D: 12542
30000 @ 10+0.1 th 1 Take 3. Contempt 12 with quadratic part. Will also test for [0, 5]
21-02-18 jh bad_bishop_1 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 4454 W: 968 L: 1079 D: 2407
sprt @ 10+0.1 th 1 Bad Bishop Detection
21-02-18 jd blockedBishops diff
LLR: -1.41 (-2.94,2.94) [0.00,5.00]
Total: 2216 W: 461 L: 513 D: 1242
sprt @ 10+0.1 th 1 Take 2. Simpler version, require low mobility instead of other conditions. Prio -2 at least until prev. test resolves.
16-02-18 II tmm diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 48216 W: 10616 L: 10590 D: 27010
sprt @ 10+0.1 th 1 My take on time management.
21-02-18 sn LSB2' diff
LLR: -2.94 (-2.94,2.94) [0.00,5.00]
Total: 9921 W: 1639 L: 1722 D: 6560
sprt @ 60+0.6 th 1 LTC: Try k+=3. Bench: 5619410
21-02-18 jd blockedBishops diff
LLR: -1.62 (-2.94,2.94) [0.00,5.00]
Total: 5135 W: 1108 L: 1155 D: 2872
sprt @ 10+0.1 th 1 Attempt to solve a few small weaknesses observed in SF losses - penalty for bishop trapped behind fixed pawn centre of its own colour. Fixed bench now, sorry.
21-02-18 sn LSB2' diff
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 29693 W: 6677 L: 6399 D: 16617
sprt @ 10+0.1 th 1 Try k+=3. Bench: 5619410
21-02-18 sg unsafe_checks diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 16320 W: 3541 L: 3592 D: 9187
sprt @ 10+0.1 th 1 Last test failed hard because i forgot to check for the rook and bishop attacks! Now the corrected version.
21-02-18 SC fb9f7abc369c1baa0b95867 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 28717 W: 6084 L: 6076 D: 16557
sprt @ 10+0.1 th 1 Does quadratic contempt gains Elo in self-play?
21-02-18 An kingpawns diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 12215 W: 2629 L: 2700 D: 6886
sprt @ 10+0.1 th 1 STC: Take 3
21-02-18 mc pinsee diff
LLR: -2.95 (-2.94,2.94) [-3.00,1.00]
Total: 15103 W: 3231 L: 3425 D: 8447
sprt @ 10+0.1 th 1 Remove pin-aware SEE: possibly useless
21-02-18 SC fb9f7abc369c1baa0b95867 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 21878 W: 4743 L: 4767 D: 12368
sprt @ 10+0.1 th 1 Does quadratic contempt gains Elo in self-play? Higher basis contempt
20-02-18 fa NOBSD diff
ELO: -1.71 +-3.8 (95%) LOS: 19.0%
Total: 14000 W: 3048 L: 3117 D: 7835
14000 @ 10+0.1 th 1 Correct Opposite Bishop Scale Down Elo Worth
20-02-18 SC fb9f7abc369c1baa0b95867 diff
ELO: 150.96 +-2.9 (95%) LOS: 100.0%
Total: 30000 W: 14358 L: 2086 D: 13556
30000 @ 10+0.1 th 1 Take 1 of trying to replace offset contempt by nonlinear one. Testing to check which variant is efficient against bad engines, will then move to [-3, 1]. See discussion on github for issues fixing in dynamic contempt case. Thx to Stefan for checking.
21-02-18 sg unsafe_checks diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 170 W: 3 L: 160 D: 7
sprt @ 10+0.1 th 1 Count preparation of discovered checks always only as unsafe check
20-02-18 SC symmetricContempt diff
ELO: 153.93 +-2.9 (95%) LOS: 100.0%
Total: 30000 W: 14556 L: 2071 D: 13373
30000 @ 10+0.1 th 1 Take 2 logarithmic (I have to fix the SF7 bench for take 1)
21-02-18 sg safe_checks diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 15055 W: 3320 L: 3377 D: 8358
sprt @ 10+0.1 th 1 Count for check bonuses also preparation of some discovered check
15-02-18 sn tweak_futility_margins diff
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 109891 W: 24566 L: 23891 D: 61434
sprt @ 10+0.1 th 1 Tweak futility margins values, and introduce an array to store them. Tested as SPRT[0..5] because of the introduction of the array, which is complicated and forces a slow memory access.
20-02-18 Fi ContemptTweak diff
ELO: 167.82 +-3.1 (95%) LOS: 100.0%
Total: 30000 W: 15828 L: 2368 D: 11804
30000 @ 10+0.1 th 1 Contempt tweak
21-02-18 Ro BishopPairMobility diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 11474 W: 2528 L: 2603 D: 6343
sprt @ 10+0.1 th 1 Take 2: smaller penalty and bonus
20-02-18 tv Razor3 diff
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 24496 W: 5470 L: 5210 D: 13816
sprt @ 10+0.1 th 1 Margin 590, I expect this to fail fast
21-02-18 jd passedSecondPush diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 14579 W: 3122 L: 3182 D: 8275
sprt @ 10+0.1 th 1 Another idea - consider second push of an unblocked passed pawn.
20-02-18 An kingpawns diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 10908 W: 2319 L: 2397 D: 6192
sprt @ 10+0.1 th 1 STC: Take 2, move to king evaluation, park running test at prio -5
20-02-18 Ro BishopPairMobility diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 8801 W: 1896 L: 1984 D: 4921
sprt @ 10+0.1 th 1 Bishop pair value correction based on duo mobiity (fied bench)
20-02-18 jd passedpawnTweak diff
LLR: -2.41 (-2.94,2.94) [0.00,5.00]
Total: 10603 W: 2269 L: 2324 D: 6010
sprt @ 10+0.1 th 1 Bench shows that on average we score passed pawns with our rook/queen behind them _less_ than passers with no rooks/queens either supporting or attacking. Try to make this more logical.
17-02-18 lb threatByPawn diff
LLR: -2.96 (-2.94,2.94) [-3.00,1.00]
Total: 107825 W: 17850 L: 18177 D: 71798
sprt @ 60+0.6 th 1 simplify ThreatBySafePawn
20-02-18 An kingpawns diff
LLR: -0.98 (-2.94,2.94) [0.00,5.00]
Total: 4000 W: 876 L: 900 D: 2224
sprt @ 10+0.1 th 1 STC: if enemy king is not on our kings file (and their neighbouring files) reduce safety for every pawn on these files which can be captured by enemy pawns from the same files
19-02-18 tv RazorEG diff
LLR: 0.52 (-2.94,2.94) [0.00,5.00]
Total: 46000 W: 10152 L: 9903 D: 25945
sprt @ 10+0.1 th 1 Try improving on the green test before LTC: < 4 * ValueRook
20-02-18 tv Razor3 diff
LLR: -2.94 (-2.94,2.94) [0.00,4.00]
Total: 65540 W: 14467 L: 14372 D: 36701
sprt @ 10+0.1 th 1 As suggested by St├ęphane try 580 margin
19-02-18 Fi UnbiasedContempt diff
ELO: 154.63 +-3.0 (95%) LOS: 100.0%
Total: 30000 W: 14970 L: 2435 D: 12595
30000 @ 10+0.1 th 1 Take 2
20-02-18 jd KnightEdgePasser diff
LLR: -2.56 (-2.94,2.94) [0.00,5.00]
Total: 13089 W: 2746 L: 2796 D: 7547
sprt @ 10+0.1 th 1 Only apply penalty in knight endgames, also count twice if there are passers on both A & H files.
18-02-18 jo remove_dyn_ct diff
LLR: -2.96 (-2.94,2.94) [-3.00,1.00]
Total: 89968 W: 19357 L: 19697 D: 50914
sprt @ 10+0.1 th 1 Remove dynamic contempt and explicitly reset 'bestValue' at the start of the MultiPV loop. Current dynamic contempt is buggy in case of a multiPV search because 'bestValue' refers to the previous PV line. (Local simplification test sprt(-4, 0) is looking good so far.)
20-02-18 sn LSB diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 39544 W: 8684 L: 8621 D: 22239
sprt @ 10+0.1 th 1 Only for doubly supported passed pawns, take 2
20-02-18 jd lessDoubled diff
LLR: -1.52 (-2.94,2.94) [0.00,5.00]
Total: 8204 W: 1785 L: 1812 D: 4607
sprt @ 10+0.1 th 1 Don't score phalanx doubled, rescale penalty to compensate based on bench.
20-02-18 sn LSB diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 22069 W: 4783 L: 4806 D: 12480
sprt @ 10+0.1 th 1 Only for doubly supported passed pawns. Bench changes at higher depths.
20-02-18 jd doubled diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 15821 W: 3444 L: 3498 D: 8879
sprt @ 10+0.1 th 1 Small penalty for doubled pawns regardless of whether the front pawn is supported.
18-02-18 mc eval_style diff
LLR: 2.95 (-2.94,2.94) [-4.00,0.00]
Total: 75666 W: 16482 L: 16616 D: 42568
sprt @ 10+0.1 th 1 Verify coding style tweaks do not introduce some hidden regression
20-02-18 pr ps_doubled diff
LLR: -1.10 (-2.94,2.94) [0.00,5.00]
Total: 2618 W: 564 L: 600 D: 1454
sprt @ 10+0.1 th 1 broader definition of "doubled" with a reduced penalty. Try #1.
19-02-18 sn LSB diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 6753 W: 1441 L: 1539 D: 3773
sprt @ 10+0.1 th 1 Take 2, half effect
19-02-18 An pl diff
LLR: -2.95 (-2.94,2.94) [-3.00,1.00]
Total: 36146 W: 7855 L: 8091 D: 20200
sprt @ 10+0.1 th 1 STC: retire piecelist and index
19-02-18 fa QT1 diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 22857 W: 4966 L: 5040 D: 12851
sprt @ 10+0.1 th 1 Try the opposite (-20)
19-02-18 sg xray_checks2 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 34458 W: 7627 L: 7588 D: 19243
sprt @ 10+0.1 th 1 Treat now bishop and rook xray checks like unsafe checks (Park with prio -1)
12-02-18 pe tunedtm diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 131831 W: 22135 L: 21867 D: 87829
sprt @ 60+0.6 th 1 LTC Tuned values with adjustment. Take 3
19-02-18 sn LSB diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 16363 W: 3584 L: 3635 D: 9144
sprt @ 10+0.1 th 1 Protected passed pawns
19-02-18 jd quadraticKingEg diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 7689 W: 1635 L: 1729 D: 4325
sprt @ 10+0.1 th 1 Try to add a small quadratic term to endgame safety.